Developing and implementing a pharmacy risk screening tool

University Hospital Ayr is an acute NHS hospital of 333 beds, serving south Ayrshire within NHS Ayrshire and Arran. The JAC Electronic Prescribing system was first introduced in the hospital in 2000, and has since been implemented throughout inpatient areas within Ayr, and subsequently, Biggart Hospitals (a geriatric outpost hospital of another 143 beds).

A well-established pharmacy department based within University Hospital Ayr provides a clinical pharmacy service to key clinical areas and further inpatient areas as departmental staffing allows. Historically, the provision of clinical pharmacy cover with high-risk, ‘core’ areas was assigned on a ward-by-ward basis by the senior team.

In early 2010, a serious clinical incident within a ward not receiving a routine/comprehensive clinical pharmacy service highlighted the deficiencies of this model of service provision. Investigations identified that harm had occurred in part because of inappropriate and extended antibiotic use, and suggested that proper clinical pharmacy involvement would have prevented patient harm. As a result of internal discussions, the idea of using the Hospital Electronic Prescribing and Medicines Administration system (HEPMA) to assist in assigning risk scores on a patient (rather than ward) basis was mooted and taken forward by the lead HEPMA pharmacist in partnership with the clinical pharmacy team.

Reducing harm from medication is a key component of the Scottish Patient Safety Programme,(1) and clinical pharmacy is a recognised defence in this area.(2) However, the nature of the patient journey within the hospital setting exposed weaknesses in provision of this service.

These deficiencies were exposed when the pre-existing, ward-based approach failed to identify a preventable medication incident. Joint work was undertaken by the clinical pharmacy and HEPMA teams to use real-time prescribing information from the HEPMA system to identify patients at the greatest risk of harm. It was hoped that such a tool would assist the pharmacy team target its services towards those patients at greatest need.

Early discussions within the clinical pharmacy team highlighted data viewed within HEPMA that influenced the priority they assigned to patients when screening a ward list. These included:

Absence of properly documented allergy status

Large numbers of unverified prescriptions (more than ten)

As such, this list was taken as a starting point for developing a scoring system for prescribing for all inpatients, with scoring resulting in patients being assigned to a low, medium or high risk category. The lead HEPMA pharmacist developed a report using the Crystal Reports XI® software to query the data held within the HEPMA system in real time to perform this scoring at scheduled intervals of the working day. The results were then emailed to all members of the clinical pharmacy team automatically for use and evaluation.

It was agreed that a period of evaluation and development of the scoring system was required, and weekly sessions were arranged between the lead HEPMA pharmacists and a suitably representative sample of clinical pharmacists (minimum of four clinical pharmacists with cover from both medical and surgical specialties). At these sessions, a random sample of five patients from each risk category were reviewed by the group and the score assigned rated (on a 0–10 range) against the groups own validation of the patient’s pharmaceutical risk. Wherever possible, specific observations were made with respect to the scoring and these were incorporated into the risk assessment algorithm for future sessions.

In addition to these formal sessions, the clinical pharmacy team used the screening tool within their day-to-day work and reported any observations to the lead HEPMA pharmacist, who made amendments to the scoring algorithm as required on an ad hoc basis. Any contentious changes were taken back to the wider clinical team for discussion and taken forward where a clear consensus of opinion existed within the clinical team.

The development over time and assessments of the accuracy of the tool can be seen in Figure 1.

The scoring of low- and high-risk categories showed a clear improvement over the 15-week evaluation and development period. The medium risk category was less clear cut, with considerable disagreement within the group over the scoring of specific non-formulary medicines. After the 15-week period, a team decision was taken to move from the intensive evaluation period into active use of the report was made after this period.

Several limitations to this method of calculating risk were identified during the development of the screening tool:

Because of this, intelligent use of the screening tool is required in conjunction with other sources of information, such as existing care plans (recorded electronically on Electronic Prescribing) and communication with ward staff.

During this evaluation and development process the scoring matrix went through 18 different iterations to reach the current version. Ongoing development with changes in medications requires ongoing development in response to requests from the clinical pharmacy team.

With the Pharmaceutical Risk screening tool now well embedded within the clinical pharmacy teams’ day to day workflow, an evaluation of it’s impact is appropriate. A crude analysis of the time interval between prescription and verification event within the HEPMA system has been performed comparing two six month periods: 1/4/2009-31/10/2009 (before implementation) and 1/4/1-31/10/11 (post implementation) for patients within University Hospital Ayr.

Comparison of the interval between prescribing and clinical pharmacy for all prescriptions within University Hospital Ayr (Table 1); suggests reductions in both the average and median interval before pharmacist intervention.

Considering the origins of this tool in trying to ensure clinical pharmacy input for high-risk medicines and medicines for specific disease states, a further two comparisons were performed to evaluate the impact on verification time for medicines targeted as requiring early clinical pharmacy assessment by the team (Table 2) and those medicines considered high risk and held on the restricted medicines list (Table 3); a similar picture is apparent here, with a particularly encouraging reduction in the median time to clinical pharmacy activity.

These results need to be viewed with some caution because of multiple confounding factors, including:

However, the results are encouraging and, taken along with positive feedback from the clinical pharmacy team, present a positive picture.

This screening tool is now well embedded within the workflow of the clinical pharmacy team, with reports covering all ward areas delivered automatically by email to all team members twice daily and on demand, as required. Working with this tool has helped the clinical team focus their efforts on the patients most in need of their attention.

In addition to ongoing developments in response to new medicines and feedback by the clinical pharmacy team, we hope to be able to incorporate data from other clinical systems into the screening tool in the near future. The potential benefits in incorporating data from laboratory systems in particular presents an exciting prospect with regard to streamlining therapeutic drug monitoring for all patients.

It should be noted that although the development of this type of screening tool is not dependent on HEPMA, the complexity of the latest iteration of the scoring algorithm is such that performing it manually for a single patient would be impractical. Furthermore, when one considers that scoring calculations are carried out for all inpatients within the Ayr and Biggart hospitals (with plans for further expansion), the value/necessity of an HEPMA system in this scenario becomes apparent.

The data provided by HEPMA systems presents exciting new avenues to develop and target existing services in addition to the more obvious statistical analysis of prescribing and administration.